So if the scientist(s) will reports that they had it enabled, it would mean that OpenAI lied, when they made a statement about not using their chats for training, right?:)